> > > SMT siblings share caches and other hardware, halt polling will
> > > degrade its sibling performance if its sibling is busy
> >
> > Do you have any real scenario benefits? As the polling nature, some
> > cloud providers will configure to their preferred balance of cpu usage
> > and performance, and other cloud providers for their NFV scenarios
> > which are more sensitive to latency are vCPU and pCPU 1:1 pin，you
> > destroy these setups.
> >
> > Wanpeng
>
Run a copy (single thread) Unixbench, with or without a busy poll program in its SMT sibling, and Unixbench score can lower 1/3 with SMT busy polling program
Can this case show this issue?
-Li
> True, it benefits for our real scenario.
>
> this patch can lower our workload compute latency in our multiple cores VM
> which vCPU and pCPU is 1:1 pin, and the workload with lots of computation and
> networking packets.
>
> -Li
On Thu, Jul 22, 2021, Li,Rongqing wrote:
> > > > SMT siblings share caches and other hardware, halt polling will
> > > > degrade its sibling performance if its sibling is busy
> > >
> > > Do you have any real scenario benefits? As the polling nature, some
> > > cloud providers will configure to their preferred balance of cpu usage
> > > and performance, and other cloud providers for their NFV scenarios
> > > which are more sensitive to latency are vCPU and pCPU 1:1 pin，you
> > > destroy these setups.
> > >
> > > Wanpeng
> >
>
>
> Run a copy (single thread) Unixbench, with or without a busy poll program in
> its SMT sibling, and Unixbench score can lower 1/3 with SMT busy polling
> program
Rather than disallowing halt-polling entirely, on x86 it should be sufficient to
simply have the hardware thread yield to its sibling(s) via PAUSE. It probably
won't get back all performance, but I would expect it to be close.
This compiles on all KVM architectures, and AFAICT the intended usage of
cpu_relax() is identical for all architectures.
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 6980dabe9df5..a07ecb3c67fb 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -3111,6 +3111,7 @@ void kvm_vcpu_block(struct kvm_vcpu *vcpu)
goto out;
}
poll_end = cur = ktime_get();
+ cpu_relax();
} while (kvm_vcpu_can_poll(cur, stop));
}
> -----邮件原件-----
> 发件人: Sean Christopherson <[email protected]>
> 发送时间: 2021年7月27日 9:26
> 收件人: Li,Rongqing <[email protected]>
> 抄送: Wanpeng Li <[email protected]>; Paolo Bonzini
> <[email protected]>; Ingo Molnar <[email protected]>; Peter Zijlstra
> <[email protected]>; kvm <[email protected]>; LKML
> <[email protected]>
> 主题: Re: 答复: [PATCH] KVM: Consider SMT idle status when halt polling
>
> Rather than disallowing halt-polling entirely, on x86 it should be sufficient to
> simply have the hardware thread yield to its sibling(s) via PAUSE. It probably
> won't get back all performance, but I would expect it to be close.
>
> This compiles on all KVM architectures, and AFAICT the intended usage of
> cpu_relax() is identical for all architectures.
>
Reasonable, thanks, I will resend it
-Li
> diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index
> 6980dabe9df5..a07ecb3c67fb 100644
> --- a/virt/kvm/kvm_main.c
> +++ b/virt/kvm/kvm_main.c
> @@ -3111,6 +3111,7 @@ void kvm_vcpu_block(struct kvm_vcpu *vcpu)
> goto out;
> }
> poll_end = cur = ktime_get();
> + cpu_relax();
> } while (kvm_vcpu_can_poll(cur, stop));
> }
>